Now that you have a working vocabulary for who populates these stories, here's where the stories themselves actually come from. Every chronicle you'll eventually read on this site — the wars, the boons, the family dramas — is pulled from one of eighteen specific books, or occasionally from the two great epics that sit alongside them. This article is about those eighteen books: what makes something a Purana at all, how they're organized, and why the same episode can read differently depending on which one you're in.
What "Purana" Means
Purana simply means "ancient" or "old" — texts concerned with the primordial past, the deep history of gods, sages, and kings. Unlike the Vedas, which you'll meet properly in the next article in this series, the Puranas are classified as smriti, "that which is remembered," rather than shruti, "that which is heard." That distinction matters practically: smriti texts have acknowledged human compilers, tolerate regional variation, and were composed — and continually re-composed, expanded, and adjusted — over many centuries, rather than being treated as fixed, authorless, eternal sound. Nearly everything this site's story series tells you comes from this smriti layer, not the stricter Vedic one.
The Five Traditional Traits
Tradition holds that a genuine Purana should contain five categories of content, known as the Pancha Lakshana, "five characteristics." Not every surviving Purana observes all five with equal attention — some lean heavily into one or two — but the definition itself tells you a great deal about what this whole genre is actually for.
| Trait | What It Covers |
|---|---|
| Sarga | The primary creation of the universe |
| Pratisarga | Secondary creation — the cycles of destruction and re-creation that follow the first |
| Vamsha | The genealogies of gods and sages |
| Manvantara | The great ages of the world, each presided over by a different Manu |
| Vamshanucharita | The detailed dynastic histories of the solar and lunar royal lines descended from those genealogies |
Look closely at the last two traits, Vamsha and Vamshanucharita — genealogy and dynastic history — and you're looking at the working definition of this entire website. FamilyRoots.in's genealogical database, tracing thousands of figures back through the Solar and Lunar dynasties, and its narrative chronicle series built from those same lineages, are both, in the most literal sense, an extension of what a Purana was always supposed to contain. This site isn't retelling the Puranas from the outside. It's continuing the specific kind of project — Vamsha and Vamshanucharita — the Puranas themselves define as central to the genre.
The Eighteen Mahapuranas
Tradition counts eighteen major Puranas, the Mahapuranas, alongside a further set of lesser Upapuranas. Several of them will already sound familiar once you start reading this site's story series, because entire chronicles draw directly from a single one.
| Purana | Traditional Length | Known For |
|---|---|---|
| Vishnu Purana | ~23,000 verses | One of the clearest, most systematic accounts of the Solar and Lunar dynasties |
| Bhagavata Purana | ~18,000 verses | Krishna's life in full, and the source for much of this site's Dashavatara material |
| Shiva Purana | ~24,000 verses | Shiva's mythology in depth — Andhaka's and Tarakasura's chronicles both draw on it |
| Markandeya Purana | ~9,000 verses | Contains the Devi Mahatmya, the core text behind this site's Three Charitas series |
| Skanda Purana | ~81,000 verses — the largest by far | Centered on Kartikeya (Skanda), Tarakasura's eventual slayer; organized into several major khandas, or sections, several times the length of any other single Purana |
| Padma Purana | ~55,000 verses | Wide-ranging cosmological and devotional material, including the traditional three-fold classification below |
| Brahma Purana | ~10,000 verses | One of the oldest surviving Puranas, heavy on regional pilgrimage material |
| Vayu Purana | ~24,000 verses | Close in content to the Brahma Purana; strong genealogical material |
| Narada Purana | ~25,000 verses | Devotional and ritual-focused, named for the sage Narada |
| Agni Purana | ~15,400 verses | An unusually broad encyclopedic text, covering everything from ritual to grammar to medicine |
| Bhavishya Purana | ~14,500 verses | Framed as prophecy of future events |
| Brahmavaivarta Purana | ~18,000 verses | Krishna and Radha-centered devotional material |
| Linga Purana | ~11,000 verses | Shiva-centered, focused on the linga as a symbol of the formless divine |
| Varaha Purana | ~24,000 verses | Centered on Vishnu's Varaha avatara |
| Vamana Purana | ~10,000 verses | Centered on Vishnu's Vamana avatara — the source for much of Mahabali's chronicle |
| Kurma Purana | ~17,000 verses | Centered on Vishnu's Kurma avatara |
| Matsya Purana | ~14,000 verses | Centered on Vishnu's Matsya avatara; also a major source for royal genealogies |
| Garuda Purana | ~19,000 verses | Best known for its detailed material on death, the afterlife, and funeral rites |
| Brahmanda Purana | ~12,000 verses | Named for the Brahmanda, the cosmic egg — heavy cosmological content |
One name credited across nearly the entire textual tradition — Vedas, epics, and all eighteen Puranas alike
Tradition credits the sage Vyasa with compiling all eighteen Mahapuranas, the same figure credited with dividing the Vedas and compiling the Mahabharata — a single author-figure standing behind nearly the entire textual tradition this site draws from. You'll meet the fuller account of Vyasa's role in the next article in this series.
The Three-Fold Classification
Later tradition, recorded within the Padma Purana itself, sorts the eighteen Mahapuranas into three groups based on the three gunas, or qualities, of nature — sattva (clarity, goodness), rajas (passion, activity), and tamas (inertia, darkness) — and links each group to which deity receives the greatest emphasis: Puranas centered on Vishnu are classed as sattvic, those centered on Brahma as rajasic, and those centered on Shiva as tamasic. This classification is traditional rather than universally accepted — plenty of later scholars and devotees, particularly within Shaiva traditions, have pushed back on a scheme that ranks Shiva-centered texts as the "lowest" of the three qualities — but it's worth knowing the framework exists, since you'll sometimes see a Purana described this way without explanation.
Why the Same Story Reads Differently Across Puranas
Because the Puranas are smriti rather than shruti, they were never expected to agree with each other word for word, or sometimes even in outline. A story like Mahishasura's death, or Tarakasura's boon, can appear with genuinely different details depending on which Purana you're reading — a different parentage, a different sequence of events, a different emphasis on which character's perspective matters most. This site's own story series follows the same principle its underlying genealogical database already commits to: textual variants are noted honestly rather than smoothed over into one artificial "correct" version. When you see a variant-box flagging a different textual tradition inside one of this site's chronicles, this is the reason it's there — not an error being corrected, but the Puranic tradition genuinely disagreeing with itself, the way smriti has always been comfortable doing.
Eighteen books, one enormous family, and no single author insisting everyone tell it exactly the same way.
How These Texts Actually Survived
It's worth knowing, before you read another chronicle on this site, that there is no single original copy of any Purana sitting in a vault somewhere. What survives is many hundreds of disagreeing handwritten copies, and everything you can actually read today — including every source cited in this site's own story series — is a scholar's careful reconstruction from among them.
Every Purana was composed in Sanskrit, but Sanskrit has no single native script of its own. It was written down wherever it traveled in whatever script the local scribe used — Sharada, Newari, Maithili, Bengali, Devanagari, Telugu, Grantha, Malayalam, and more, depending on the region. Before any of it was written at all, this material was carried orally for generations, and once it finally was committed to writing, it was recorded on palm-leaf and birch-bark manuscripts — materials that physically decay within a few generations in India's climate. Every manuscript had to be periodically recopied by hand simply to survive, and every recopying was a chance for a scribal slip, a regional retelling, or a deliberate expansion to enter the text. This is the real, physical reason behind the kind of textual variance this article keeps flagging — not carelessness, but the literal mechanics of how a handwritten tradition stays alive across a thousand years.
Every surviving Purana passed through hands like these, over and over, for a thousand years
The Skanda Purana, already noted above as the longest of the eighteen, is also the most dramatic case study in this survival story. The oldest dated Puranic manuscript known anywhere is a Nepalese palm-leaf copy of the Skanda Purana from the year 810 CE — but the text as most people encountered it in printed editions had drifted so far from that early form that scholars long considered its authentic original essentially lost. An international critical-edition project, anchored at Leiden University, has spent decades since the 1990s working back from surviving manuscript fragments across multiple countries to reconstruct something closer to the earliest recoverable version — work that is still ongoing.
Other Puranas have their own dedicated reconstruction projects: the Oriental Institute at the Maharaja Sayajirao University of Baroda produced the standard critical edition of the Vishnu Purana from forty-three manuscripts spanning northern and southern Indian scripts; the All-India Kashiraj Trust in Varanasi has focused specifically on the Puranas named for Vishnu's avataras — Varaha, Kurma, Vamana — sourcing manuscripts from institutions as far apart as Pune and Harvard University; and the Bhandarkar Oriental Research Institute in Pune, best known for its critical edition of the Mahabharata, remains a central reference point across nearly every Puranic text-critical project.
The physical manuscripts themselves are scattered wherever history left them — Indian university and temple libraries, the national archives of Nepal, and collections built up in Europe and the United States during the colonial period, when Sanskrit manuscripts were actively collected by foreign scholars and institutions. Today, the largest coordinated effort to document and preserve what remains is run by the Government of India through the National Mission for Manuscripts, established in 2003 and substantially expanded in 2025 under the name Gyan Bharatam Mission. It has documented several million manuscripts across the country out of an estimated ten million believed to exist, digitized hundreds of thousands of them into a national digital archive, and built a network of over a hundred dedicated resource and conservation centres — with Sanskrit material making up roughly seven in every ten manuscripts the mission has catalogued.
No original text survives anywhere. What survives is a thousand years of careful hands, copying, disagreeing, and copying again — and what you read today is the best current attempt to listen through all of them at once.
Puranas Versus Itihasa
One more distinction worth having early: the Ramayana and the Mahabharata are not, technically, Puranas at all. They belong to a separate category, Itihasa, "thus indeed it was" — epic history, rather than the cosmological and genealogical sweep of a Purana. In practice the two categories overlap constantly in how they're read and retold, and this site's own name, FamilyRoots' Itihasa-Puranic Lineages, holds both words side by side deliberately, because its story series draws on both categories without much concern for the technical boundary between them. But if you ever wonder why Ravana's story sits differently in tone from Tarakasura's, part of the answer is that one comes primarily from Itihasa and the other primarily from Purana — two related but distinct genres, doing related but distinct kinds of work.
The Bhagavad Gita is not a Purana, and it isn't a standalone book in its own right either — technically, it's 700 verses embedded within a single stretch of the Mahabharata's Bhishma Parva, the dialogue between Krishna and Arjuna occurring just before the Kurukshetra war begins. By the strict genre lines this article has been drawing, that makes it Itihasa, the same category as the rest of the Mahabharata around it.
In practice, though, the Gita has been treated as something closer to shruti than ordinary Itihasa ever is. Alongside the Principal Upanishads and the Brahma Sutras, it forms the Prasthanatrayi, the three-fold canonical foundation every major school of Vedanta philosophy — Shankara's, Ramanuja's, Madhva's, and others — has had to write a formal commentary on to be taken seriously as a philosophical system at all. That's an extraordinary status for a text that is, by strict genre classification, just one chapter of a much longer epic: it's sometimes called the Gitopanishad, "the Gita as Upanishad," precisely because of how far its authority has outgrown its technical home inside the Mahabharata.
You now know what a Purana actually is — a smriti text built around the Pancha Lakshana, especially the genealogical Vamsha and Vamshanucharita material this entire site is an extension of — which of the eighteen Mahapuranas feeds which of this site's story series, why variant tellings across them aren't errors but the tradition's own comfort with multiple versions, and how a thousand years of hand-copied palm-leaf manuscripts is the physical reason those variants exist at all. The next article in this series, The Vedas, Briefly, goes one layer further back — to the shruti material older and stricter than anything in this article, and to Vyasa's own fuller story as the one figure organizing nearly all of it.
References
Primary Sources
- Vishnu Purana, Book III (the Pancha Lakshana and the traditional list of eighteen Mahapuranas)
- Padma Purana, Uttara Khanda (the sattva/rajas/tamas three-fold classification)
- Bhagavata Purana, Skandha XII (a parallel account of the eighteen Puranas and their traditional lengths)
- M.M. Pathak (ed.), Critical Edition of the Vishnu Purana, Oriental Institute, Maharaja Sayajirao University of Baroda, 1997–99
- R. Adriaensen, H.T. Bakker, H. Isaacson (eds.), The Skandapurana, Volume I, Leiden critical edition project, Egbert Forsten, 1998
- Government of India, Ministry of Culture — National Mission for Manuscripts / Gyan Bharatam Mission, public documentation figures as of 2025
- Bhagavad Gita, Bhishma Parva of the Mahabharata, chapters 23–40 (or 25–42 in some editions' numbering); traditional commentaries of Shankara, Ramanuja, and Madhva on the Prasthanatrayi
The three-fold sattva/rajas/tamas classification is itself found within only one Purana (the Padma) rather than being independently attested everywhere, and is treated by many later commentators as a sectarian framing favoring Vaishnava texts rather than a neutral, universally accepted scheme. It's presented here as a traditional framework worth knowing, not an endorsed ranking. The verse counts given for each Purana are the traditional figures most commonly cited (several are recorded within the Puranas' own self-references); actual surviving manuscripts vary, sometimes considerably, and should be treated as an approximate sense of relative scale rather than an exact, source-verified count.